https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78264
--- Comment #11 from Eric Botcazou <ebotcazou at gcc dot gnu.org> --- > thanks for fixing this. Shouldn't we keep the PR open, though? It's > still an ice-on-invalid. Given that it didn't raise much interest as ice-on-valid-code, I wouldn't expect it to raise interest as ice-on-invalid at all. :-) More seriously, the ICE is not a regression and only enabled with checking: if (TREE_CODE (expr) == INTEGER_CST) { if (operand_equal_p (expr, boolean_true_node, 0)) return noexcept_true_spec; else { gcc_checking_assert (operand_equal_p (expr, boolean_false_node, 0)); return noexcept_false_spec; } } so I wouldn't bother about it.